Performance testing of an axial flow fan in an A-frame test facility

(1)

April 2019 Supervisor: Dr J. Muiyser Co-supervisor: Dr D.N.J Els

Thesis presented in partial fulfilment of the requirements for the degree of Master of Mechanical Engineering in the Faculty of Engineering at

Stellenbosch University by

(2)

Declaration

By submitting this thesis electronically, I declare that the entirety of the work contained therein is my own, original work, that I am the sole author thereof (save to the extent explicitly otherwise stated), that reproduction and publication thereof by Stellenbosch University will not infringe any third party rights and that I have not previously in its entirety or in part submitted it for obtaining any qualification. Date: April 2019

(3)

Plagiaatverklaring / Plagiarism Declaration

1 Plagiaat is die oorneem en gebruik van die idees, materiaal en ander

intellektuele eiendom van ander persone asof dit jou eie werk is.

Plagiarism is the use of ideas, material and other intellectual property of another’s work and to present is as my own.

2 Ek erken dat die pleeg van plagiaat 'n strafbare oortreding is aangesien dit ‘n vorm van diefstal is.

I agree that plagiarism is a punishable offence because it constitutes theft.

3 Ek verstaan ook dat direkte vertalings plagiaat is.

I also understand that direct translations are plagiarism.

4 Dienooreenkomstig is alle aanhalings en bydraes vanuit enige bron

(ingesluit die internet) volledig verwys (erken). Ek erken dat die woordelikse aanhaal van teks sonder aanhalingstekens (selfs al word die bron volledig erken) plagiaat is.

Accordingly all quotations and contributions from any source whatsoever (including the internet) have been cited fully. I understand that the reproduction of text without quotation marks (even when the source is cited) is plagiarism.

5 Ek verklaar dat die werk in hierdie skryfstuk vervat, behalwe waar anders aangedui, my eie oorspronklike werk is en dat ek dit nie vantevore in die geheel of gedeeltelik ingehandig het vir bepunting in hierdie

module/werkstuk of ‘n ander module/werkstuk nie.

I declare that the work contained in this assignment, except otherwise stated, is my original work and that I have not previously (in its entirety or in part) submitted it for grading in this module/assignment or another

module/assignment.

J van Eck

(4)

Abstract

Performance testing of an axial flow fan in an A-frame test facility The primary objective of this research and experimental investigation is to determine the performance of an axial flow fan in an A-frame test facility. Commissioning of the A-frame test facility required certain modifications which allowed the facility to be used for testing purposes. Modifications include a new fan drive system, fan bridge, heat exchanger and structural supports. The bell mouth roundness was improved upon, achieving a more constant tip clearance around the casing, and all air leaks around the facility were sealed to prevent loss of mass flow. Tests were conducted on the B2-fan, according to standardised fan testing procedures in ISO 5801 (2007) Type-A, to obtain the fan characteristics at selected blade angles. The measurements recorded at the A-frame test facility included outlet velocity, inlet velocity, fan torque and rotational speed measurements. The outlet velocity profiles differ for the two heat exchanger outlets, but the total volumetric flow rate differs only by 3% for the two sides. It was also found that the velocity profile stays similar with a change in blade angle, but increases in magnitude as the blade angle is increased. Pressure measurements according to ISO 5802 (2008), at the walls of the plenum chamber, were regarded as unusable due to it being much lower than the expected pressure.

With the use of the system draft equation from Kröger (2004, 8.1.30) it was possible to calculate a recovery coefficient for the A-frame plenum chamber at different blade angle settings of the B2-fan, with values of 0.505, 0.685 and 0.412 for the 28º, 31º and 34º blade angles respectively. The recovery coefficient calculated for the M-fan equalled 0.143 and 0.322 for the 34º and 35º blade angles respectively. This value represents the portion of kinetic energy that is converted back into a pressure. The sensitivity of the recovery coefficient to change in volumetric flow rate and heat exchanger loss coefficient was also investigated with the use of a Monte Carlo simulation. The importance of accurately measuring the volumetric flow rate is shown by a normal distribution of the recovery value with the use of the mean and standard deviation of the volumetric flow rate. A functional A-frame test facility can now be used to test different heat exchanger and fan configurations. By changing the fan, heat exchanger type or area, the performance of the A-frame setup can be analysed for various heat exchanger and fan configurations.

(5)

Uittreksel

Toets van ‘n aksiaalwaaier in ‘n A-raam toets fasiliteit

Die primêre doelwit van hierdie navorsing en experimentele ondersoek is om die werking van ‘n aksiaal waaier binne ‘n A-raam toets fasiliteit te bepaal. Die ingebruikstelling van die A-raam toetsfasiliteit het sekere aanpassings benodig wat die fasiliteit bruikbaar vir toets doeleindes gestel het. Aanpassings sluit ‘n nuwe waaier aandryfstelsel, waaier brug, warmte uitruiler en strukturele versterkings in. Die klokmond inlaat se rondheid is verbeter wat daartoe gelei het dat ‘n meer konstante waaier lempunt spasiëring binne die omhulsel bereik is. Verder is alle lekplekke verseel sodat daar geen massavloei verlore gaan nie. Toetse is gedoen op die B2-waaier volgens die standard waaier toets prosedures soos in ISO 5801 (2007) Tipe A om die waaier karakteristieke by sekere lem hoeke te verkry. Die metings wat by die A-raam fasilitiet geneem is sluit die uitlaat snelheid, inlaat snelheid, waaier wringkrag en die waaier omwentelingspoed in. Die uitlaat snelheidsprofiel verskil vir die twee warmte-uitruilers, maar die totale volumetriese vloeitempo tussen die twee kante slegs met 3%. Die snelheidsprofiel bly naastenby dieselfde vir verskillende lemhoeke en verander slegs met ‘n konstante waarde regoor die profiel, Druklesings is ook geneem in die vier mure volgens ISO 5802 (2008), maar die resultate daarvan as nutteloos beskou weens dit heeltemal te lae resultate gegee het.

Deur gebruik te maak van die sisteem vloei vergelyking, vanaf Kröger (2004, 8.1.30), is die druk herwinningskoëffisiënt vir die B2-waaier in die

A-raam opstelling bereken om 0.505, 0.685 en 0.412 vir die 28º, 31º and 34º lemhoeke respektiewelik te wees. Die herwinningskoëffisiënt vir die M-waaier is bereken om 0.143 en 0.322 vir die 34º en 35º lemhoeke respektiewelik te wees. Hierdie koëffisiënt word gebruik om die gedeelte kinetiese energie wat in druk omgeskakel word binne die plenum kamer te beskryf. Die sensitiwiteit van hierdie koëffisiënt vir verandering in die volumetriese vloeitempo en warmte uitruiler druk verlies koëffisiënt is ook bereken deur gebruik te maak van die Monte Carlo metode. Die belangrikheid van akkurate metings om die volumetriese vloeitempo te bereken word duidelik gemaak deur die normaal verspreiding van die herwinnings koëffisiënt deur gebruik te maak van die gemiddelde en standaardafwyking van die volumetriese vloeitempo. ‘n Funsionele A-raam toetsfasiliteit kan nou gebruik word om verskillende warmte uitruiler en waaier samestellings mee te toets. Deur die waaier, tipe warmte uitruiler of die area van die warmte uitruiler te verander, kan die A-raam opstelling getoets word vir verskillende warmte uitruiler en waaier opstelling senarios.

(6)

Acknowledgements

I would like to acknowledge my friends and family who stood by me through the good and the not so good times. To my mom, Naomi van Eck, I would like to say thank you for always being there and for all your support. You are a winner and can do anything you set your mind to. My dad, Johannes Jacobus van Eck, I would like to thank for your inspiration and motivation. I always enjoy our long conversations about new patents and projects entering the global market. My sister Petro van Eck, thank you for the example you set for all of us. You are indeed a driven person who gets the job done! My brother, Jacobus van Eck, I would like to thank for your humour in almost every situation. You always find a way to make someone laugh with the strangest of stunts. Jana-Mari, I would like to say thank you for always being there to light up my day when I needed it the most. Also thank you for your time assisting me in performing the tests at the test facilities. Without you it would have taken much longer. To Jana-Mari’s family I say thank you for your interest in the progress and all your support.

I would also like to thank Hein Joubert for his intellectual contributions when I got stuck and needed some good insight in design matters. I appreciate your friendship. Many thanks to Dr. Jacques Muiyser and Dr. Danie Els for the opportunity to do this research project. I am very grateful for all your input and guidance throughout the whole project. The project wouldn’t have been such a success without you. Dr. Francois Louw, I would like to thank you and your colleagues at Kelvion for all your effort and time regarding the finned heat exchanger tubes. Also, many thanks to Kelvion for the sponsoring of the tubes at a discounted price.

I would like to thank the staff of the workshop for their contributions in the manufacturing process, our long conversations and especially to Maurisha who I could always bother for some more tools. Calvin Harmse and Jacobus Samuels, I enjoyed working together with you. Your humour is always there to cheer someone up. Many thanks also to Ferdie and Cobus Zietsman for their guidance regarding the design process. Nathi and Julian, thank you for always assisting when needed. Finally I would like to thank my Lord and Saviour, Jesus Christ, for the opportunities we have. I am grateful even for the difficult times of growth.

(7)

List of figures

Page

Figure 1.1: A-frame plenum ACC ... 1

Figure 2.1: Schematic ISO 5801 (2007) fan testing facility ... 4

Figure 2.2: Section of an A-frame array (Kröger, 2004: 5.6.12) ... 6

Figure 2.3 Entrance flow behaviour at a thin sheet bundle (Kröger, 2004: 5.6.2) .. 7

Figure 2.4: Two row plate finned heat exchanger loss factor(Kröger, 2004: 5.6.10) ... 7

Figure 2.5: Fan and plenum chamber test setup (Meyer and Kröger, 1998) ... 8

Figure 2.6: Operating point due to Krec (Meyer, 1998) ... 10

Figure 2.7: Effect of tip clearance on fan performance for flow rates above 6 m3/s (Wilkinson and Van der Spuy, 2015) ... 12

Figure 2.8: A-frame test facility constructed by Böck (2017) ... 14

Figure 2.9: B2-fan ... 15

Figure 2.10: M-fan schematic (Wilkinson, 2017) ... 16

Figure 3.1: Initial fan drive system Böck (2017) ... 18

Figure 3.2: Bevel gearbox concept ... 19

Figure 3.3: Belt drive concept ... 19

Figure 3.4: Motor mount welded assembly design ... 20

Figure 3.5: Belt cover and motor mount ... 20

Figure 3.6: Manufactured and installed fan bridge ... 21

Figure 3.7: Machined bearing mounting surface ... 22

Figure 3.8: Perforated plates Böck (2017) ... 23

Figure 3.9: Heat exchanger tubes: (a) Rectangular plate fin; (b) Round finned; (c) Wavy finned flattened tube ... 24

Figure 3.10: Analytical finned heat exchanger tube bundle pressure drop ... 26

Figure 3.11: Induced draft wind tunnel from Kröger (2004, 5.2.2) ... 27

Figure 3.12: Finned tube heat exchanger test section ... 28

Figure 3.13: Analytical and experimental bundle pressure drop ... 28

Figure 3.14: Heat exchanger calibration with multiple netting configurations ... 29

Figure 3.15: Heat exchanger assembly design ... 30

(10)

Figure 3.17: Velocity traverse measurement grid (dimensions in mm) ... 31

Figure 3.18: Vertical inflow velocity measurement ... 33

Figure 3.19: Vertical inflow velocity measurement results ... 34

Figure 3.20: Horizontal approach velocity measurements ... 35

Figure 3.21: Horizontal approach velocity around bell mouth perimeter ... 35

Figure 3.22: Outlet velocity profiles in at 31º blade angle without end plates ... 36

Figure 3.23: Outlet velocity profiles in at 31º blade angle with end plates ... 36

Figure 3.24: Pressure measurement points ... 37

Figure 4.1: Fan static pressure rise at different blade angle settings ... 41

Figure 4.2: Fan power consumption at different blade angle settings ... 41

Figure 4.3: Fan static efficiency at different blade angle settings ... 42

Figure 4.4: B2-fan outlet velocity profiles in at 28º blade angle ... 44

Figure 4.7: M-fan outlet velocity profiles in at 34º blade angle ... 45

Figure 4.10: A-frame ACC and ISO 5801 (2007) B2-fan power consumption .... 47

Figure 4.11: A-frame ACC and ISO 5801 (2007) M-fan power consumption... 48

Figure 5.1: A-frame ACC schematic representation ... 50

Figure 5.2: ACC supports and cylindrical flow area ... 52

Figure 5.3: Coefficient of drag for two dimensional bodies (Kröger, 2004) ... 52

Figure 5.4 Contraction ratio for round tubes and parallel plates (Kröger, 2004) .. 55

Figure 5.5: Heat exchanger loss coefficient for normal flow Khe ... 56

Figure 5.6: Fan static pressure rise coefficient KFs for B2-fan ... 57

Figure 5.7: Fan static pressure rise coefficient KFs for M-fan ... 58

Figure 5.8: Volumetric flow rate normal distribution curve for B2-fan ... 63

Figure 5.9: Fan static pressure rise coefficient 95% confidence intervals for the B2-fan ... 65

Figure 5.10: Normal distribution of Krec from variability in volumetric flow rate for the B2-fan ... 66

(11)

Figure 5.12: Normal distribution of Krec from variability in KFs for the B2-fan ... 68

Figure 5.13: Recovery normal distribution curve for B2-fan ... 69

Figure 5.14: Recovery normal distribution curve for M-fan ... 69

Figure B.1: Load cell calibration setup ... 80

Figure B.2: Load cell calibration ... 81

Figure B.3: Schematic torque transducer calibration setup ... 82

Figure B.4: Resulting torque at voltage output ... 83

Figure B.5: Torque transducer calibration setup ... 83

Figure B.6: Induced draft wind tunnel pressure transducers ... 85

Figure B.7: Wind tunnel upstream pressure transducer calibration ... 85

Figure B.8: Wind tunnel nozzle pressure transducer calibration ... 86

Figure B.9: Anemometer calibration inside induced draft wind tunnel ... 87

Figure B.10: Anemometer A3 calibration curve ... 88

Figure B.11: Torque transducer calibration setup ... 90

Figure B.12: Pressure transducer calibration ... 90

Figure B.13: Shaft rotational speed calibration ... 91

Figure C.1: A-frame supports loss coefficient ... 93

Figure C.2: Fan bridge loss coefficient ... 94

Figure D.1: Fan static pressure rise tests at different blade angle settings ... 96

Figure D.2: Fan power consumption tests at different blade angle settings ... 97

(12)

List of tables

Page

Table 2.1: B2-fan Characteristics ... 15

Table 2.2: M-fan characteristics ... 16

Table 3.1: Expected operating points ... 30

Table 3.2: Plenum chamber pressure measurements ... 37

Table 4.1: Fan power consumption and average volumetric flow rate ... 43

Table 4.2: Heat exchanger volumetric flow rates ... 46

Table 5.1: Heat exchanger parameters ... 55

Table 5.2: B2-fan static pressure rise coefficient ... 58

Table 5.3: M-fan static pressure rise coefficient ... 59

Table 5.4: Krec change as a result selected parameters ... 61

Table 5.5: Average and standard deviation of Khe during calibration ... 64

Table B.1 : A-frame ACC facility torque transducer results ... 84

Table B.2: Anemometer calibration results ... 87

Table B.3: Anemometer recalibration results ... 89

Table C.1: Vertical supports loss coefficients ... 92

Table C.2: Fan bridge loss coefficients ... 93

Table C.3: Heat exchanger loss coefficient ... 94

Table C.4: Fan static pressure rise coefficient ... 95

Table D.1: B2-fan tests at 28º blade angle according to ISO 5801 (2007) ... 98

Table D.4: Eastern side outlet velocity measurements at 28º blade angle ... 99

Table D.5: Western side outlet velocity measurements at 28º blade angle ... 100

Table D.7: Western side outlet velocity measurements at 31º blade angle ... 100

(13)

Nomenclature

Abbreviations

ACC Air Cooled Condenser Variables

A Area [m2]

CD Coefficient of drag

Cp Specific heat at constant pressure [J/kgK]

Cv Specific heat at constant volume [J/kgK]

D Diameter [m] d Diameter [m] E Voltage [V] F Force [N] G Mass velocity [kg/m2s] L Length [m]

𝑚̇ Mass flow rate [kg/s]

m mass [kg] N Rotational velocity [rpm] p Pressure [N/m2_] P Power [W] r Radius [m] T Temperature [ºC] T Torque [N.m] v Velocity [m/s]

V Volumetric flow rate [m3/s]

V Voltage [V]

W Width [m]

Greek Symbols

β Thermal expansion coefficient η Efficiency µ Viscosity ρ Density ε Effectiveness Subscripts a Air abs Absolute amb Ambient app Approach b Bridge

c Fan cross sectional area/Casing cal Calibrated

(14)

D Drag

do Downstream

e Effective fan flow area

F Fan f Fin fr Frontal h Hub he Heat exchanger l Lognitudinal meas Measured n Nozzle o Outer r Root rec Recovery s Static sc Settling chamber sh Shaft sup Support t Total/Transverse tr Tube rows up Upstream Constant Values

(15)

1 Introduction

1.1 Background

Fans are used in various industrial processes where cooling, mixing and ventilation is required. These fans may differ in shape and size depending on the particular application. Factors such as noise, cost and performance characteristics are considered when selecting the correct fan to be used for the specific application. For high flow rate at low pressure, axial flow fans are used. The volumetric flow rate of these fans at a specified rotational speed increases with an increase in the number of fan blades (Kröger, 2004: 6.0.1). Increasing the angle of attack or rotational speed also results in an increase in the volumetric flow rate.

In many industries cooling of a working fluid is required. Dry cooling is normally used where water is scarce. This requires a closed loop system where the working fluid is cooled and reused in the next cycle. Air-cooled condensers find application in many of these industries and are used in induced draft and forced draft configurations. Forced draft dry cooling normally consists of an axial flow fan which forces air flow through the heat exchanger bundles. The air-cooled condenser, ACC, may take on various shapes, but commonly used today is the A-frame plenum as shown in Figure 1.1.

(16)

For the induced draft air-cooled condenser, the fan sucks air through the heat exchanger with the fan blowing air from the plenum to the atmosphere. An advantage of such a system is the reduced recirculation of hot plume air due to a higher exit velocity. A disadvantage is that the fan is blowing hotter and thus thinner air than the forced draft ACC’s.

The forced draft A-frame plenum assists with effective drainage of the condensate, minimizing the facility ground surface area and decreasing the length of the distribution steam ducts. Condenser tubes are angled downwards from the steam header to increase the area of condenser surface exposed to the airflow. The axial flow fans located below the plenum blows air into the plenum chamber which is then forced through the condenser.

1.2 Problem definition

Meyer and Kröger (1998) conducted tests on different fan and heat exchanger combinations in order to determine the influence that these combinations have on the losses inside of the plenum chamber. Tests were conducted at a BS 848 (1980) facility where the total flow rate and the outlet velocity profile was known. Heat exchangers were mounted so the flow would enter and exit normally. The recovery coefficient Krec introduced by Meyer and Kröger (1998), indicates that adding a plenum chamber with a heat exchanger to the fan outlet, results in a portion of the kinetic energy of the air being converted into pressure.

The value of the recovery coefficient is of importance as Meyer and Kröger (1998) indicates that the operating point of a fan installed in a plenum with a heat exchanger may not necessarily lie on the point where the fan pressure rise curve crosses the system pressure drop curve. Investigating if a difference in performance, for fans installed in a standardised fan test facility and fans installed in an A-frame ACC, can be quantified by a recovery coefficient can be of importance for the design of ACC systems. Meyer and Kröger (1998) found that there exists a critical minimum distance between the heat exchanger and the fan outlet. Should this distance be further decreased, it would result in the recovery value Krec decreasing radically. There thus exist certain design parameters which would should be considered to achieve a certain value for Krec.

1.3 Aim and objectives of the study

The aim of this study is to determine if a difference in performance between a fan installed in a standardised fan testing facility as tested according to ISO 5801 (2007) and a fan installed in an A-frame ACC setup can be quantified by a recovery coefficient. Furthermore, the accuracy of recovery due to measurement accuracy should also be investigated.

(17)

The following objectives were identified to obtain the required data for an investigation on the fan performance results from a standard fan testing facility and an A-frame ACC setup:

1. The available 1:6 scale A-frame fan test facility had to be modified and commissioned for testing purposes to ensure that testing can be done safely without excessive vibration.

2. Measure the performance of a fan in a standardised fan test facility according to ISO 5801 (2007) and also in an A-frame test facility.

3. Investigate how the accuracy of measurements influences the accuracy to which a result can be represented.

(18)

2 Literature survey

A study of the available literature was needed to obtain the required knowledge of the working of ACC’s and the current methods being used for testing the already installed fans. This chapter aims to provide the reader with a brief overview of the current literature available regarding the testing of already installed ACC fans.

2.1 Standardised fan performance testing

Fans are tested in standardised fan test facilities to determine the performance characteristics they have. For axial flow fans the BS 848 and even more recently updated ISO 5801 (2007) are used. Different installation types are recognized of which Type-A in Figure 2.1, which employs a free inlet and free outlet, is used and can be modified to perform type B, C and D installations.

Figure 2.1: Schematic ISO 5801 (2007) fan testing facility

The inlet bell mouth is located at 1, followed by 2, which is a set of louvres acting as a flow resistance to change the system operating point. A set of flow straighteners at 3 follows the louvres to even out abnormalities in the flow velocity profile. An auxiliary fan at 4 overcomes frictional effects at higher flow rates after which another set of flow straighteners removes the swirl caused by the auxiliary fan. Guide vanes at 5 are used to more evenly distribute the flow into the settling chamber at 6 where wire mesh screens at 7 are used to improve the flow inside the chamber. The fan to be tested is located at 8 inside a bell mouth with a machined cylindrical section to achieve a constant fan blade tip clearance. The different installation types are listed below:

1. Type A – free inlet and free outlet. 2. Type B – free inlet and ducted outlet. 3. Type C – ducted inlet and free outlet. 4. Type D – ducted inlet and ducted outlet.

(19)

Le Roux (2010) investigated the method and accuracy of simulating axial flow fans with computational fluid dynamics. Tests were performed at the standard Type-A test facility and the results compared with the simulated model. When testing the B2-fan, designed by Bruneau (1994), Le Roux (2010) compared his fan performance curves to those of Stinnes (1998). Le Roux (2010) found similar results but, Stinnes (1998) measured better performance characteristics due to smaller tip clearances achieved. Stinnes (1998) managed the smaller tip clearances due to the shroud of the facility being fixed to the frame of the drive shaft, preventing them to move independently. Le Roux (2010) performed tests at a 3 mm tip clearance where Stinnes (1998) was able to test between 1 mm and 1.5 mm. Le Roux (2010) found that his final model over predicts the tested fan static pressure rise at the design flow rate of 16 m3/s by 1.4%.

Louw (2015) investigated the flow field in the vicinity of an axial flow fan operating at low flow rates. Tests were conducted on the B2-fan according to the BS 848 standard in a Type-A test facility. A new model was developed for the actuator-disk method. The new model gave a reasonable representation of the experimental results at the design flow rate, but still deviated from experimental results at low flow rates. It was however an improvement on the existing model.

Augustyn (2013) tested the performance of the L1-, L2- and N-fans, which are scale models of existing industrial fans, according to the BS 848 standards in a Type-A standardised test facility. These were scale fan tests of actual ACC fans. Augustyn (2013) found that at high flow rates the L1-fan and the L2-fan followed the trend of the full scale model. A noticeable difference between the L2-fan and the full scale model fan pressures could be the result of a difference in blade angles said Augustyn (2013). The N-fan also showed a comparable trend. It was however found that the L2-fan power correlated very well with the full scale L-fan. Furthermore a difference of 10-15% between the full scale and tested fan efficiencies was found. Augustyn (2013) also tested and demonstrated the validity of the scaling laws for the L2- and N-fan at different rotational speeds.

Wilkinson and Van der Spuy (2015) performed fan tests on the B2-fan to investigate the effect of fan tip clearance and fan tip modifications on the performance of the fan. These tests were also performed in the same facility as used by Le Roux (2010), Augustyn (2013) and Louw (2015). It was found that tip clearance contributes significantly to the fan performance. Their blade tip modifications also showed improvement of fan performance on larger tip clearances.

2.2 Performance of a fan in a plenum chamber

The performance of a fan inside of a plenum chamber may be influenced by various contributing factors such as geometry and system losses. Experiments completed regarding the performance of a fan in a plenum chamber is investigated to gain an understanding of the effect of the plenum chamber on the fan performance.

(20)

2.2.1 Plenum chamber and system losses

Experiments conducted by Meyer and Kröger (1998) indicated that the inlet geometry of the heat exchanger could significantly affect the performance characteristics of the plenum chamber. Obstacles located upstream and downstream from the fan are added as loss coeficients in the system by Kröger (2004, 8.1.30). These would include support, upstream, bridge, heat exchanger, downstream, jetting and outlet losses. Flow that apporaches the heat exchanger at an angle which is not perpendicular will add an additional loss term to the system. This is the case for A-frame plenums where the flow which also has a vertical component from the fan, is turned by the heat exchanger fins. Figure 2.2 shows a schematic representation of the flow through the heat exchangers of a section of an A-frame array. The jetting and outlet losses represent the losses due to the flow turning and exiting the array as a result of the symmetry plane where the adjacent heat exchanger bundles directs the flow towards each other at an angle.

Figure 2.2: Section of an A-frame array (Kröger, 2004: 5.6.12)

Flow that approaches and exits a heat exchanger non-perpendicularly has higher losses due to separation of the flow as it meets the fins at an angle. This separation region acts as a flow constriction. Figure 2.3 shows how flow that approaches a finned heat exchanger at an angle separates upon entry as it is forced to turn suddenly and re-attaches to recover static pressure, thereby not dissipating the complete velocity component perpendicular to the fins as would be the case in a porous medium.

(21)

Figure 2.3 Entrance flow behaviour at a thin sheet bundle (Kröger, 2004: 5.6.2) From Figure 2.4 it can be seen that flow that approaches and exits a finned heat exchanger bundle perpendicularly, or at 0º, has a lower pressure loss across the bundle than in the case of flow approaching the heat exchanger bundle at 30º and leaving perpendicularly, or at 0º. These tests were performed by placing a plate finned heat exchanger in wind tunnel and changing the approach- and exit angle of the flow with respect to the heat exchanger fins. A loss factor was then calculated at different volumetric flow rates through the bundle.

(22)

2.2.2 Fan and plenum chamber tests

Meyer (1996) constructed a delta shaped ACC with a casing diameter of 1.542 m for use with the B-fan. An industrial type heat exchanger was used. The outlet velocity profile was measured at a plane, not parallel to the heat exchanger outlet surface, but rather at the outlet plane which would normally be above the top steam header. These measurements were possible due to symmetry planes added to the setup to represent a single A-frame ACC as if it was installed in an A-frame array. Meyer and Kröger (1998) installed various different heat exchanger-plenum configurations over the outlet side of the Type-A standard fan test facility to test the effect that different heat exchanger and plenum chamber geometries have on the flow losses inside a plenum chamber. Figure 2.5 shows a schematic of the test setup.

Figure 2.5: Fan and plenum chamber test setup (Meyer and Kröger, 1998)

Adding a plenum chamber with a heat exchanger to a fan, forces the flow to slow down and spread out across the heat exchanger area. The higher the flow resistance of the heat exchanger, the more evenly the flow will be distributed over the outlet surface. As the flow is slowed down the velocity component of the air is converted into pressure which is then recovered inside of the plenum chamber.

Testing at this standardised test facility allowed for the volumetric flow rate through the system to be accurately measured with the use of the calibrated bell mouth. The volumetric flow rate through the facility could also be changed with the use of a throttling device.

(23)

When using equation 2.1 from Kröger (2004, 6.4.4), it is assumed that all of the kinetic energy is lost inside of the plenum chamber. This is however not the case as shown by Meyer and Kröger (1998).

∆𝑝_𝐹𝑠 = 𝐾_ℎ𝑒𝜌𝑣_ℎ𝑒2⁄ + 𝛼2 _𝑒𝜌𝑣_ℎ𝑒2⁄ (2.1) 2 Where Khe and αe represents the heat exchanger loss coefficient and velocity distribution correction factor respectively. The velocity through the heat exchanger

vhe is calculated with the use of the conservative volumetric flow rate at the point where the fan static pressure rise curve intersects the system pressure loss curve with the use of equation 2.2.

𝑣ℎ𝑒 = 𝑉/𝐴𝑓𝑟 (2.2)

The conservative volumetric flow rate V is divided by the frontal area of the heat exchanger Afr to calculate the velocity vhe through the heat exchanger. Recovery introduced by Meyer and Kröger (1998), is a factor which describes the portion of kinetic energy of the air that is converted into pressure inside of the plenum chamber. This suggests that not all of the kinetic energy is lost inside of the plenum chamber. This results in the fans performing at a higher pressure in situ than in the standard fan test facilities at the same volumetric flow rate. The implication thereof is that the system operating point would not necessarily lie on the fan curve as determined by the standard fan testing procedures.

Meyer and Kröger (2004) found that the maximum Krec value does not necessarily coincide with the maximum fan static efficiency. This is an indication that the value for recovery is not necessarily constant for the fan at different operating points. Figure 2.6 shows how the operating point would differ from the point where the fan curve crosses the system pressure loss curve according to Meyer and Kröger (1998). It was also found that the value for Krec increased as the volume flow rate increased past the maximum fan static efficiency. From Kröger (2004, 6.4.4), equation 2.3 is used to calculate the value for Krec in a simplified forced draft heat exchanger where the only losses are that of the heat exchanger. The kinetic energy correction factor depicted by αe accounts for the losses not accounted for when assuming a uniform outlet velocity profile.

∆𝑝_𝐹𝑠+ 𝐾_𝑟𝑒𝑐𝜌𝑣_𝐹2_{⁄ = 𝐾}₂

(24)

Figure 2.6: Operating point due to Krec (Meyer, 1998)

According to Kröger (2004, 6.4.4) recovery of the kinetic energy results in the

volumetric flow rate being higher than the predicted value where the fan and system curves intersect. This can be an important aspect to consider when designing ACC systems. A higher volumetric flow rate than expected provides an additional safety factor with the increased cooling capability. Not taking this into account can result in over designing of an ACC system which leads to large, possibly unnecessary, extra manufacturing costs.

Meyer and Kröger (2004) performed a numerical investigation on the effect of the B-fan performance on the aerodynamic behaviour of the plenum chamber. Various different blade angles were tested and the results showed recovery values ranging from 0.4 to 0.8. Engelbrecht (2018) performed a numerical investigation of the performance of the B2a-fan in a forced draft ACC. A recovery coefficient of 0.527 was found through the numerical simulation. Engelbrecht (2018) validated his findings by comparing an experimental design with the numerically calculated recovery. A good correlation was found with the results form Meyer and Kröger (1998).

2.3 Effect of distorted inflow

The axial flow fans, as mentioned in section 1.1, are subject to operating conditions when installed on site, which differ from the ideal testing conditions for which they are designed and tested in the standard fan testing facilities. This can be a result of changes in ambient conditions such as temperature, humidity and wind. A small scale ACC fan installed indoors can be subject to distorted inflow over the fan due to structural components and spatial constraints negatively affecting the airflow towards the bell mouth. The effect of distorted inflow on fan performance is thus an important consideration due to the A-frame ACC being installed indoors with limited space and some obstructions that may negatively affect the flow. Salta and

(25)

Kröger (1995) investigated the effect of reducing the height of the fans from ground level. It was found that it reduced the efficiency of the edge fans due to a higher cross flow at the fan inlet. Fourie et al. (2015) investigated the effect of wind on an ACC by lowering the platform height in a small scale test facility. It was found that the volumetric effectiveness decreases with a decrease in platform height as the fans further away from the edge fan induces a flow across the edge fan inlet bell mouth. This leads to flow separation at the edge fan and a decrease in the volumetric effectiveness. Salta and Kröger (1995) found that reducing the platform height of a single or multiple fan row ACC reduces the volumetric efficiency exponentially. Kröger (2004, 6.4.7) said that the absence of a bell inlet reduces the fan performance. This is due to separation at the edge of the inlet to the fan. Duvenhage et al. (1996) showed that replacing a well-rounded bell mouth inlet with a conical or cylindrical inlet will decrease the flow rate by about 3 and 5 percent respectively. Even if a well-rounded bell mouth is fitted, other factors such as the surrounding walls and the height of the fan to the floor surface can also adversely affect the fan performance. Monroe (1979) found that the horizontal approach velocity should not exceed one half of the axial flow velocity through the fan. Walls at different distances around the test facility can result in mal-distributed flow across the fan due to different approach velocities around the bell mouth circumference, resulting in an uneven load condition experienced by the fan blades (Muiyser et al., 2014). Van Rooyen (2007) completed a study on the effect that wind has on the performance of an ACC. It was found that the fans situated on the perimeter of the ACC were affected the most and the fans inwards were mostly unaffected by the wind. It was also found that the static pressure at the edge fans was unequally distributed across the fan inlet. For this study wind would not have an influence on the test results as the facility is installed indoors.

Duvenhave and Kröger (1996) numerically investigated the influence of wind on the forced draft ACC performance and found that cross winds mainly lead to a reduction in upstream fan volumetric performance. Owen (2010) also performed a numerical investigation on the performance of an ACC under windy conditions with the use of data recorded at a full scale ACC. Wind velocity measurements at different elevations above ground level were used as input parameters to the numerical model. The ambient temperature measurements were used to determine the air temperature at the platform. Owen (2010) found that a good correlation was achieved between the numerical model and the test data. Discrepancies in the results obtained, although being small, can be described by the weather conditions and numerical inaccuracies from the numerical model used. Owen (2010) found that the wind screens appeared to have a positive effect on the fans that were located directly upstream of the screens. However, it seemed to be reducing the downstream fan performance.

(26)

2.4 Effect of fan tip clearance on performance

Venter and Kröger (1992) developed correlations for the V-type, eight bladed fan that showed the effect that increasing the fan tip clearance has on the flow rate and fan static pressure. It was shown that an increase in tip clearance results in a linear decrease in flow rate and static pressure respectively. This corresponds with Stinnes (1998) who achieved better performance characteristics with the B2-fan than Le Roux (2010) due to Stinnes (1998) being able to test as smaller tip clearances. Wilkinson and Van der Spuy (2015) performed tip clearance tests on the B2-fan. The same type of linear relationship is seen for the B2-fan where an increase in tip clearance results in a linear decrease in flow rate and static pressure rise as seen in Figure 2.7.

Figure 2.7: Effect of tip clearance on fan performance for flow rates above 6 m3/s (Wilkinson and Van der Spuy, 2015)

Tip modification tests were also done by Wilkinson and Van der Spuy (2015) to determine the effect on fan performance. The S20 end plate tip modification, which extended 20 mm from the centre of the leading edge to the pressure side with a 4 mm tip clearance, resulted in the same static efficiency as the 2 mm tip clearance. This can be of importance in full scale applications where it is difficult to achieve a constant and small tip clearance on the large diameter fans and fan casings.

(27)

2.5 In situ volumetric air flow measurements

According to ISO 5802 (2008) the volumetric flow rate in an airway may be determined by measuring the velocity at various points and calculating the mean velocity. By using the velocity area method the volumetric flow rate can then be calculated from the velocity profile. Another method is by using a pressure differential measurement device like an orifice plate, Venturi or a flow nozzle. Muiyser (2012) investigated the use of different air flow sensors in order to establish the method for measuring the volumetric flow rate through the ACC cooling fans. The use of a propeller anemometer, pressure probe, and also a two axis ultrasonic anemometer was considered. Due to its slower response time and accurate measurement up to 40º off from its main axis Muiyser (2012) did not use pressure probes to measure the flow through the cowling. Muiyser (2012) used both propeller and ultrasonic anemometers to measure the inlet flow velocities. The ultrasonic anemometers were used to measure the inlet flow and the propeller anemometers used to measure the out of plane flow component inside the fan casing. Maulbetsch et al. (2011) suspended propeller anemometers under the fans to monitor the inlet velocity at selected fans. These measurements were made to determine the effect of wind on the ACC performance.

Venter (1990) made use of a five holed pressure probe to measure the velocity components, making use of a spherical coordinate system, at the fan inlet and fan outlet respectively. To measure the velocity profile at the heat exchanger outlet, Venter (1990) made use of a vane anemometer which was traversed at a distance of 25 mm from the outlet of the heat exchanger surface. The outlet velocity profile of the unit was measured between the apexes of the A-frames with the use of the same vane anemometer. Meyer and Kröger (1998) measured the outlet velocity profile of a heat exchanger by traversing a 1.9 m rod and eight anemometers that were evenly spaced. Zapke (1997) used an aluminium beam fitted with 10 anemometers to perform a full scale traverse of the outlet side an A-frame ACC. These anemometers were spaced 200 mm away from the outer side of the heat exchanger bundle.

2.6 In-situ fan performance measurements

When measuring the pressure rise across a fan a certain test length of the duct is required at which measurements shall be taken according to ISO 5802 (2008). For measurements upstream of the fan, the pressure measurement plane shall not be closer than 1.5 diameters De and for downstream measurements no closer than 5

diameters De. Where De is the equivalent diameter of a non-circular cross section.

ISO 5802 (2008) stipulates that these distances may be shorter if the flow is already stable at a closer distance. Some fans are directly connected to the plenum chamber on either the inlet or outlet sides of the fan. When taking the pressure measurements it should be done in a plane as close as possible to the face of the plenum chamber to which the fan is connected. These points of pressure measurements should be located in areas with no significant air velocity. When the pressure measurements

(28)

between the pressure taps in the walls differs by less than 5%, they can be interconnected to achieve an average gauge pressure (ISO 5802, 2008).

2.7 A-frame test facility

Böck (2017) designed and constructed an A-frame test facility to investigate the performance of the B2-fan with a forced draft configuration. Bruneau (1994) designed two experimental fans, the B1- and B2-fan, which were tested according to the BS 848 Type A standards. The requirements of the design for Böck (2017) included the following:

1. A forced draft design with an A-frame plenum and apex angle of 60º. 2. Use of the existing B2-fan to compare results with standardised tests

performed on the fan in a Type A fan testing facility. 3. Minimum flow distortion at the bell mouth inlet.

Figure 2.8 shows the constructed test facility. Böck (2017) had a few limitations to consider during the design which included the width and height of the facility. The width was limited to 3.5 m and the height to 5.2 m due to the service trenches and the maximum hook height of the crane respectively. The motor, which was a reconditioned 15 kW three phase motor, was too heavy and big to be put on the fan bridge and would have affected the flow inside of the plenum chamber negatively. It was thus mounted below the fan on the inlet side to the bell mouth.

(29)

The constructed facility was designed on a 1:6 scale. Perforated plates and guide fins were constructed and tested by Böck (2017). The purpose of these plates was to enable the system to be throttled and tested at different operating points. Guide fins on the inside of the perforated plates were supposed to direct the flow before exiting the plenum chamber. For safety during testing, safety screens were added below the inlet bell mouth. These were to protect the facility operator during testing. A variable speed drive was used by Böck (2017) to operate the fan at the calibrated rotational speed. A potentiometer and display allow the user to set the motor to the desired rotational speed.

2.8 Selected fans for testing

Two different fans were selected for testing in the A-frame ACC test facility. They were chosen to be different to each other with regards to the hub to tip ratio, static pressure rise and operating rotational speed. This would allow the outlet velocity profiles to be compared to establish to which degree the fan influences the outlet velocity profile of the A-frame ACC.

2.8.1 B2-fan

The B2-fan designed by Bruneau (1994) utilises the high lift characteristics of the NASA LS series profile with a free vortex design and can be seen in Figure 2.9. This fan has a hub diameter Dh of 0.6 m and a total diameter Df of 1.53 m and is

used inside a casing of 1.542 m diameter. The tip clearance can be decreased or increased for safe operation inside the particular bell mouth. Table 2.1 shows the B2-fan characteristics.

Figure 2.9: B2-fan

Table 2.1: B2-fan Characteristics

Diameter 1.542 m

Number of blades 8

Hub to tip ratio 0.39

(30)

2.8.2 M-fan

The M-fan, designed by Wilkinson (2017), was designed to achieve a high total-to-static efficiency at a selected operating point. With the recent trends of using high volumetric flow rate at a low pressure rise, this fan was designed. Figure 2.10 shows a schematic representation of the M-fan and Table 2.2 shows the fan characteristics.

Figure 2.10: M-fan schematic (Wilkinson, 2017)

Table 2.2: M-fan characteristics

Diameter 1.542 m

Number of blades 8

Hub to tip ratio 0.29

Fan rotational speed 722 rpm

2.9 Summary of findings

From this literature it can be seen that adding a plenum chamber to a fan results in a portion of the kinetic energy to be converted into pressure. This recovery value has been measured by experiments by Meyer (1996) and Meyer and Kröger (1998), and more recently by Engelbrecht (2018) with the use of a numerical model. This indicates that a recovery value does exist explaining a portion of kinetic energy converted into pressure.

The tests performed by Meyer (1996) and Meyer and Kröger (1998) were however done with the use of a standardised fan testing facility allowing for accurate measurement of the volumetric flow rate through the fan which will not be the case for this experiment. Measuring of the volumetric flow rate through the A-frame test facility will be done with the use of outlet velocity measurements at the heat exchanger outlet surface. This will be the case as it is not possible to attatch a standardised fan testing facility to the inlet of a full scale ACC bell mouth inlet.

(31)

This project will develop a method to be used at full scale A-frame ACC’s to measure the recovery at full scale setups. The results of the measurements performed can then be validated by a numerical model as the one presented by Engelbrecht in a subsequent study.

(32)

3 Facility modification and commissioning

As stipulated in section 1.3 it was identified that certain modifications to the ACC facility constructed by Böck (2017) had to be made in order for it to be used for testing purposes. This chapter aims to provide an overview of the identified areas of re-evaluation, redesign and also the manufacturing and installation procedures followed.

3.1 Fan drive system

When running the fan, excessive vibration limited the rotational speed at which the fan could safely be run. Vibration caused by the fan drive system had to be addressed as it was not possible to operate the facility at the required 750 rpm without excessive structural vibration. This was possibly caused by the misalignment of the motor and the bearings on the shafts of the torque transducer as a result of too many bearings that were placed in a single line.

The initial drive system had the motor mounted close to the floor as shown in Figure 3.1. Above the motor were three separate in-line shafts, each located by two bearings, which had to be aligned with each other and the motor shaft in order to have the drive running smoothly. This in itself would have proved difficult to achieve without reference surfaces to locate the bearings on to.

Figure 3.1: Initial fan drive system Böck (2017)

When the couplings between the shafts were removed, it could be seen that the shafts were not in proper alignment. As such, this was identified as one of the reasons for the vibration when running the facility. Marks on the inside of the fan casing indicated that the blades were scraping during testing. This had to be addressed as blades striking the casing wall can lead to damage of the fan and in worst case scenarios complete destruction of the test fan and possible injuries.

(33)

Having more than two bearings on a single shaft results in residual forces acting on the bearings if they are not properly aligned. It was found necessary to minimize the amount of bearings that were required in a straight line. Two different drive systems were considered to replace the existing one.

3.1.1 Bevel gearbox concept

A bevel gearbox was considered as it would allow the motor to be fixed outside of the plenum chamber, preventing it from obstructing the flow. The relative simple sealing between the shaft and the plenum chamber with the use of a radial lip seal was also considered. This would also allow for the hub of the fan to be bolted to the bevel gearbox output shaft. However, this would require inspection of the motor bearings and possible replacement due to the installed bearings having been made specifically for vertical mounting of the motor. Figure 3.2 shows the bevel gearbox concept.

Figure 3.2: Bevel gearbox concept

3.1.2 Belt drive concept

A belt drive was also considered. It would involve the redesign of the fan bridge to accommodate the bearings for locating the shafts above and below the torque transducer. The bottom shaft would also be the shaft on which the fan would be mounted inside of the casing. With a 15 kW motor that had to be reused, it was important to install it in a location that would least adversely impact the flow inside of the plenum chamber. Mounting the motor inside the plenum was considered, but no suitable location could be identified that would not affect the air flow negatively. Due to its size, the motor would have to be mounted outside of the plenum chamber. This would lead to holes in the plenum chamber in order to accommodate the belts from the motor pulley to the pulley on the top shaft of the fan bridge. Long belts would allow for the motor to be at a fixed position outside of the plenum chamber as seen in Figure 3.3.

(34)

Three belts, two pulleys and taper lock bushings would be the main components required along with a belt cover. This was considered to be the more economic option than the bevel gearbox and also the chosen concept for the project.

In order to mount the motor on the outside of the plenum chamber, a mount had to be designed and manufactured for mounting onto the two I-beams passing on the Northern side below the plenum chamber. The original motor mount was modified and reused for this application as it would save on costs and time. Slots were made on the bottom of the motor mount to allow for belt tensioning by moving the motor mount relative to the I-beams in the Northerly direction. Diagonal supports were welded to the structure to withstand the moment acting on the mount due to the belts. Rubbers were inserted to act as absorbers of high frequency vibrations from the motor.

Figure 3.4: Motor mount welded assembly design

In order to prevent air from leaking from the plenum chamber where the belts exit and enter the plenum, a cover that seals on top of the motor and against the plenum chamber was designed. The manufactured and installed cover is shown in Figure 3.5.

(35)

Adhesive sponge was placed on the perimeter top side of the motor and also where the cover is attached to the plenum. This achieved an air seal, preventing leakage from the plenum chamber.

3.2 Fan bridge

Redesigning the fan drive system to accommodate a belt and pulley drive required the redesign of the fan bridge. The fan bridge, which locates the fan within the bell mouth, had to be designed to be stiff enough to withstand the tension of the belts

causing a moment on the bridge as seen in Figure 3.6. Rectangular tubing, 100 mm × 50 mm, which was initially oriented with its long side horizontally, was

rotated through 90º to allow for a stiffer bridge in the vertical direction.

Figure 3.6: Manufactured and installed fan bridge

The surface, to which the bell mouth is bolted and which also makes up the bottom surface of the plenum chamber, is made up of eight sections which are bolted together. In order to firmly bolt the fan bridge to the structure and more accurately locate the bridge, it was decided that it would be better to cut out the sections where the fan bridge would support itself. These sections were removed two at a time, the necessary cuts made, and then replaced. This allowed the fan bridge to be bolted directly to the I-beams that provide the structural support.

In order to mark the exact locations for the drilling of the mounting holes it was decided to insert the whole fan bridge and centre it according to the bell mouth. The holes were then marked and drilled using a magnetic drill. The welded fan bridge was lowered into position with the overhead crane and bolted to the I-beams with rubbers to act as absorbers of high frequency vibrations. Slots in the bottom of the welded assembly allowed for adjustments to be made in the longitudinal direction of the bridge as shown in Figure 3.6.

(36)

The bearing mounting surface was placed in a milling machine to achieve a flat reference face. Self-aligning bearings locating the two shafts above and below the toque transducer were located on the surface by adjustment bolts located on either side of the bearing housings. With slots also in the bearing mounting surface, the bearings could be well aligned and the shafts also moved in the East-West direction as shown in Figure 3.7. On either side of each bearing housing a threaded hole with a bolt for adjusting the location of the bearings on the reference surface was also added. This ability to move the shaft in either of the four directions enabled the fan shaft to be centred inside of the bell mouth casing.

Figure 3.7: Machined bearing mounting surface

The original couplings, as seen in Figure 3.6, were flexible stainless steel couplings locating the torque transducer between the two shafts. These however resulted in excessive vibration to the torque transducer under higher toque. This was a result of the couplings acting like springs under torsion as the torque was increased. New couplings were ordered that would be more rigid whilst still allowing for some degree of misalignment. These couplings were fixed to the shafts using taper lock bushing.

The 40 mm diameter shafts were used as the bearings only allowed a shaft diameter of 40 mm as used by Böck (2017). This is only 2 mm less than that of the standardised Type-A fan test facility. The A-frame facility fan shaft also has an additional advantage of only torque and axial load being applied to it where the standardised fan test facility has a constant moment acting on the shaft due to the mass of the fan from being installed on a horizontal shaft. It was thus concluded

(37)

that the 40 mm shaft diameter will be more than sufficient for the setup. Appendix A.1 shows a sample calculation for the fan shaft strength.

The bridge total width equals 500 mm but is only comprised of two 50 mm wide beams. Engelbrecht (2018) modelled the bridge width as 2 m. With this model being a sixth scale of the full scale model, the total bridge width should equal 0.333 m. It was however decided that having wider bridge supports will result in a much more rigid setup as safety is of great importance when testing these fans at 750 rpm. This allows for a walkway of any width to be added to the bridge and the effect thereof tested for different scenarios.

The pulley diameters were selected to be a 1:1 ratio as the motor maximum rotational speed equalled 960 rpm where the required speed was 750 rpm. Selecting the belt and pulley sizes were done according to calculations that can be seen in Appendix A.2.

3.3 Heat exchanger bundles

Heat exchanger bundles are used to increase the air contact area to cool down the low pressure steam coming from the turbines. These heat exchanger bundles normally consists of finned surfaces increasing the heat transfer rate from the steam to the surrounding air.

3.3.1 Shortcomings of existing system

The perforated plates and guide vanes used by Böck (2017) as seen in Figure 3.8 were considered to be reused for the commissioning of the small scale A-frame ACC test facility.

Figure 3.8: Perforated plates Böck (2017)

These perforated plates however only allowed for a total of 32 velocity measurement points, 16 on each side, due to each perforated plate only allowing for a single anemometer to measure the outflow of air. This was thought to be too

(38)

coarse a measurement of the flow exiting the plenum chamber. Böck (2017) suggested that the ability of the perforated plates and guide vanes to turn the flow also be investigated. It was decided that the existing perforated plates and guide vanes had to be replaced with an alternative which would better direct the flow with finer spaced guide vanes. The reason for redirecting the flow was to achieve a more accurate representation of a full scale A-frame ACC. Adding of a heat exchanger across the whole surface would also allow finer measurement capabilities of the outlet velocity profile due to no obstructed sections across the outlet surface. 3.3.2 Selection and theoretical analysis of heat exchanger tubes

Different heat exchanger tube types were considered for the application and are shown in Figure 3.9. An elliptical tube with a rectangular plate fin as shown in Figure 3.9(a) is used in ACC applications and is normally used with two or more rows. A round finned tube as shown in Figure 3.9(b) was also considered. These finned tubes are widely used today and were selected as an alternative as they are readily available. They are normally used with two or more tube rows. A wavy-finned flattened tube consists of a tube which is flattened to which wavy fins are bonded as shown in Figure 3.9(c). These heat exchanger tubes are particularly suited for ACC’s as two or more rows of elliptically or round finned tubes may be replaced by a single row of wavy finned tubes (Kröger, 2004: 5.1.8).

Figure 3.9: Heat exchanger tubes: (a) Rectangular plate fin; (b) Round finned; (c) Wavy finned flattened tube

After considering the availability, cost, delivery time and the required purpose of the heat different exchanger tubes, it was decided that the round finned heat exchanger tubes would be suitable to use for the heat exchanger bundles.

Böck (2017) designed the A-frame ACC with a bundle slope angle of 60º relative to the horizontal. Kröger (2004, 1.2.3) said that large ACC may have finned tube bundles which are sloped up to 60º. A staggered finned tube arrangement was selected as most of the literature in Kröger (2004, 5.4.17) utilises staggered finned tube arrangements for better mixing of the air. The required number of tube rows for the A-frame ACC is calculated from Kröger (2004, 5.5.11).

(39)

With a total width of 2 m and a total height of 1.5 m, the final design width and height for each heat exchanger came down to 1.8415 m and 1.4 m respectively. The parameters are listed:

Fin diameter df = 0.0572 m

Fin root diameter dr = 0.0276 m

Fin thickness tf = 0.0005 m

Fin pitch Pf = 0.0028 m

Tube outer diameter do = 0.0254 m

Arrangement Staggered

Tubes per row ntr = 29

Transverse tube pitch Pt = 0.0635 m

Longitudinal tube pitch Pl = 0.055 m

Finned tube length Lt = 1.400 m

Flow area width w = 1.8415 m Flow area height h = 1.4 m Frontal area Afr = 5.1562 m2

Minimum flow area Ac = 2.4954 m2

Ambient temperature Ta = 20 ºC

Dynamic viscosity µT = 20 ºC = 1.81454×10-5 kg/m.s

Equation 3.1 is used to calculate the minimum flow area through the finned tube heat exchanger bundle.

𝐴_𝑐 = 𝐴_𝑓𝑟− 𝑛_𝑡𝑟𝐿_𝑡[𝑑_𝑓𝑡_𝑓+ (𝑃_𝑓− 𝑡_𝑓)𝑑_𝑟] 𝑃⁄ _𝑓 (3.1) The total frontal area Afr would equal

𝐴_𝑓𝑟 = 2𝑤ℎ (3.2)

The Reynolds number is

𝑅_𝑒 =𝐺𝑐𝑑𝑟

𝜇 (3.3)

Where the value for Gc is

𝐺_𝑐 = 𝑚̇𝑎

𝐴𝑐 (3.4)

Finally the pressure drop across the heat exchanger finned tube bundle is ∆𝑝 = 𝐺𝑐2×18.93×𝑛𝑟×𝑅𝑒−0.316 𝜌 ( 𝑃𝑡 𝑑𝑟) −0.927 (𝑃𝑡 𝑃𝑑) 0.515 (3.5)

(40)

Substituting different numbers of rows into nr gives different pressure loss values

for the bundle at a constant volumetric flow rate. Figure 3.10 shows different pressure loss curves for two, four, and six row finned tube heat exchanger bundles.

Figure 3.10: Analytical finned heat exchanger tube bundle pressure drop These curves represent the analytical model for the heat exchanger bundles for flow entering and exiting the heat exchanger bundle perpendicularly. This is however not the case for the A-frame heat exchanger setup as the flow has to turn and then exit the heat exchanger bundle perpendicularly. Figure 2.4 showed how the turning of flow before entering the heat exchanger increases the total heat exchanger loss coefficient. It was however decided that the heat exchanger, being the biggest flow resistance of the A-frame ACC system, will give a reasonable indication of the number of finned tube rows required with the analytical normal flow calculations from Figure 3.10.

Wilkinson and Van der Spuy (2015) completed fan tests on the B2-fan. Tests were conducted at the design angle of 31º. From these tests the highest efficiency at a blade angle of 31º was 60.3% at a flow rate of 13.14 m3_{/s. The fan static pressure}

(41)

maximum efficiency of the fan at the design blade angle, the point where the system resistance curve cross the fan static pressure rise curve should be at 13.14 m3/s. From Figure 3.10 it was clear that not even 6 finned tube rows would provide enough flow resistance for this small scale ACC A-frame. Having more than two finned tube rows would be too heavy for the existing structure and an alternative design was required.

With the main function of the finned heat exchanger tubes being to turn the flow and provide a flow resistance, it was decided that an additional flow resistance should be placed between or in front of the heat exchanger tubes to provide the required pressure drop. Various woven steel meshes were explored as additional flow resistances. However, the pressure loss across the steel meshes were analytically calculated and it was found that it would not provide the required flow resistance.

Shade netting, which was readily available and inexpensive, was also an alternative solution as an additional flow resistance. Various shade netting ranging from 40%- up to 60% covered area were explored as a possible additional flow resistances. 3.3.3 Experimental verification of heat exchanger bundle selection

An induced draft wind tunnel shown in Figure 3.11 and described by Kröger (2004, 5.2.1) was used to measure the volumetric flow rate through and the

pressure drop across the heat exchanger bundle designed for the test section as shown in Figure 3.12. A sample calculation on the working of the wind tunnel is shown in Appendix B.1.

(42)

Figure 3.12: Finned tube heat exchanger test section

An additional pressure transducer which was calibrated according to the same procedure as described in 4.2.1, was used for measuring the static to static pressure drop across the heat exchanger bundle. The other two pressure transducers used in calculating the flow rate through the test section were also calibrated again. The calibration procedure of the pressure transducers is listed in Appendix B.2. In Figure 3.13 the pressure drop across the finned heat exchanger bundle test section is shown at various flow rates.